Action-Space Shaping for LLM Agents: Measuring and Mitigating Tool-Schema Bias
cs.AI
Submitted: 2026-09-28
Updated: 2026-09-28
Code: https://github.com/williamLyh/ToolSchemaSpace
Sources
- Granite-Function Calling Model: Introducing Function Calling Abilities via Multi-task Learning of Granular Tasks
- $\tau^2$-Bench: Evaluating Conversational Agents in a Dual-Control Environment
- From Trainee to Trainer: LLM-Designed Training Environment for RL with Multi-Agent Reasoning
- ACEBench: Who Wins the Match Point in Tool Usage?
- The Llama 3 Herd of Models
- Gemma 4 Technical Report
- GLM-5: from Vibe Coding to Agentic Engineering
- MetaTool Benchmark for Large Language Models: Deciding Whether to Use Tools and Which to Use
- ToolACE: Winning the Points of LLM Function Calling
- APIGen: Automated Pipeline for Generating Verifiable and Diverse Function-Calling Datasets
- ToolSandbox: A Stateful, Conversational, Interactive Evaluation Benchmark for LLM Tool Use Capabilities
- Gorilla: Large Language Model Connected with Massive APIs
- Qwen2.5 Technical Report
- Proximal Policy Optimization Algorithms
- Quantifying Language Models' Sensitivity to Spurious Features in Prompt Design or: How I learned to start worrying about prompt formatting
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism
- Joint Optimization of Tool Creation and Use for Large Language Model Agents
- Qwen3 Technical Report
- AppWorld: A Controllable World of Apps and People for Benchmarking Interactive Coding Agents
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems